Can an algorithm improve the design of your organization?
I believe the answer is yes.
In fact, together with two colleagues (Kim Soldal and Tore Christiansen), I am developing a software tool for this purpose. It is called Re:config.
The goal is not to automate the entire organization design process – that is not realistic (nor desirable).
What we are developing is a decision support tool: A tool that can help practitioners (leaders, internal advisors, and external consultants) create more effective organizational models.
We have created a video that explains the basic concept (it lasts for about 3 minutes):
As the video explains, what Re:config does is to help group elements (roles, teams, or sub-units) in the best possible manner.
A grouping task is like this: Should role A be organized in the same team as role B? To answer this question, one needs to know what tasks A and B have, and their interdependencies (how they are related).
This becomes a hard task when the number of elements increases.
As an example, if you have 6 roles and need to allocate them to two sub-units, there are actually 41 different ways of doing this (!).
Which of the 41 options is the most effective one? An algorithm will find the answer faster than the human brain.
By making better decisions about grouping, you will reduce coordination costs: Less time will be used in meetings and on email to coordinate with others, and more time will be spent on the core tasks.
It will also enable you to create a more flexible and modular organization, one where each sub-unit has a clear and focused mandate, with responsibility for a complete deliverable (or a set of closely related tasks).
* * *
What we have now is a prototype. It works, but like all prototypes, it will need to be further developed and testet.
For this reason we are seeking 3-4 partners (industrial companies or consulting firms) that are interested in testing the tool in their organization.
So let me know if your firm might be interested in collaborating with us on this.
P.S. The main assumptions behind the tool are described in chapter 5 in the second edition of my book.